4 sources healthy · latest episode 9 Nov 2026
descriptive claim

Current reinforcement-learning-trained models differ from models that existed many years ago.

Tim Scarfe · AI 2040: Plan A report - Daniel Kokotajlo & Thomas Larsen · 8 Sep 2026

In their words

The transcript segments this claim was taken from

“I think that these new RL trained models we have, they're not the same as what we had many years ago.”
1:28:33Report a misattribution
0.90how strongly it was put, not whether it is true
0.75importance to the conversation
0.97confidence the claim was read correctly